The Implementation of K-Means dan K-Medoids Algorithm for Customer Segmentation on E-commerce Data Transactions

  • Siagian R
  • Sirait P
  • Halim A
N/ACitations
Citations of this article
79Readers
Mendeley users who have this article in their library.

Abstract

Abstrak Data transaksi e-commerce yang semakin banyak dapat dimanfaatkan perusahaan untuk memberikan informasi yang baru. Data transaksi tersebut dapat mengungkap tentang segmentasi atau kelompok pelanggan berdasarkan kesamaan karakteristik dan perilaku pelanggan masing-masing. Salah satu teknik yang dapat digunakan untuk mensegmentasi pelanggan adalah Data Mining dengan menggunakan metode clustering. Tujuan penelitian ini adalah menerapkan metode clustering pada data transaksi e-commerce menggunakan algoritma K-Means dan K-Medoids. Hasil penelitian menunjukkan bahwa algoritma K-Means dan K-Medoids sama-sama menunjukkan hasil cluster optimal k = 3. Hasil tersebut juga sesuai dengan hasil metode elbow dan uji validitas Davies Bouldin Index yang menunjukkan bahwa jumlah cluster optimal adalah 3 (tiga). Hasil pengujian menunjukkan K-Medoids memiliki performa terbaik dengan nilai ration sebesar 0,337575 dibandingkan K-Means 0,3380724, sehingga K-Medoids digunakan dalam clustering data.sebagai cluster optimal. Hasil segmentasi pelanggan sesuai Customer Loyalty Matrix terdiri dari core customer, new customers, dan lost customer. Kata kunci: Segmentasi Pelanggan, Clustering, K-Means, Model LRFM. Abstract Nowadays, e-commerce data transactions are commonly used by companies to provide new information. The data transaction can reveal customer segmentation or groups based on the similar characteristics and behavior of each customer. Data Mining is one of technique to conduct the customer segmentation through clustering method. The study aims to applied the clustering method on e-commerce data transactions by using both K-Means and K-Medoids algorithm. The result shows that both algorithms reveal optimum of cluster result with value of k = 3. The results are also indicating the conformity with the elbow method's results and the Davies Bouldin Index validity test which shows that the optimal number of clusters is 3. The test results show that K-Medoids has the best performance with a ration value of 0.337575 compared to K-Means 0.3380724. Hence, K-Medoids are used in data clustering as the optimal cluster. The results of customer segmentation according to the Customer Loyalty Matrix consist of core customers, new customers, and lost customers.

Cite

CITATION STYLE

APA

Siagian, R., Sirait, P., & Halim, A. (2022). The Implementation of K-Means dan K-Medoids Algorithm for Customer Segmentation on E-commerce Data Transactions. SISTEMASI, 11(2), 260. https://doi.org/10.32520/stmsi.v11i2.1337

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free